Feature detection in images, videos, and/or other graphics information elements may be useful in a wide variety of contexts, to detect features such as faces, persons, objects, landmarks, landscape features, optical effects, and/or other visual features. Conventional techniques for feature detection may utilize an integral image algorithm implemented using an integral pixel value array to facilitate the feature detection process. According to such conventional techniques, use of such an integral image algorithm may include repeatedly downscaling a graphics information element and recomputing the integral pixel value array following each downscaling operation. However, the computational cost associated with performing such downscaling operations on the graphics information element to be analyzed may be significant, resulting in reduced efficiency and speed. As such, techniques for employing an integral image process without requiring repeated downscaling of the graphics information element may be desirable.